PALEOBOTANY
Southeast Asian Dipterocarp origin
and diversification driven by Africa-India
floristic interchange
Mahi Bansal^1 †, Robert J. Morley2,3†, Shivaprakash K. Nagaraju^4 †, Suryendu Dutta^5 ,
Ashish Kumar Mishra^1 , Jeyakumar Selveraj^1 , Sumit Kumar^5 , Deepti Niyolia^5 ,
Sachin Medigeshi Harish6,7, Omer Babiker Abdelrahim^8 , Shaa eldin Hasan^8 ,
Bramasamdura Rangana Ramesh^9 , Selvadurai Dayanandan6,7, Harsanti P. Morley^2 ,
Peter S. Ashton^10 , Vandana Prasad^1 *†
The evolution and diversification of ancient megathermal angiosperm lineages with Africa-India origins in
Asian tropical forests is poorly understood because of the lack of reliable fossils. Our palaeobiogeographical
analysis of pollen fossils from Africa and India combined with molecular data and fossil amber records
suggest a tropical-African origin of Dipterocarpaceae during the mid-Cretaceous and its dispersal to India
during the Late Maastrichtian and Paleocene, leading to range expansion of aseasonal dipterocarps on
the Indian Plate. The India-Asia collision further facilitated the dispersal of dipterocarps from India to similar
climatic zones in Southeast Asia, which supports their out-of-India migration. The dispersal pathway
suggested for Dipterocarpaceae may provide a framework for an alternative biogeographic hypothesis for
several megathermal angiosperm families that are presently widely distributed in Southeast Asia.
M
any of the megathermal angiosperm
families that constitute major com-
ponents of today’s tropical rainforest,
such as Annonaceae, Combreteaceae,
Ebenaceae, Myristicaceae, etc., show
disjunct pantropical distributions. The colo-
nization and diversification of megathermal
angiosperms in Southeast Asian tropical
forests is believed to be influenced by their
ancientorigininAfricaandout-of-Indiadis-
persal ( 1 – 4 ). However, the limited number of
molecular studies and the sparse fossil record
of megathermal angiosperms from the Cre-
taceous of Africa ( 1 ) and the Paleogene of
India ( 4 , 5 ) provide insufficient information
to conclude their ancient African origin or
that India aided their dispersal and diversi-
fication in Asian tropical forests. Here, we
present fossil pollen data from the family
Dipterocarpaceae that suggest their African
origin during the mid-Cretaceous and sub-
sequent dispersal to Southeast Asia.
Dipterocarpaceae are a pantropical, obligate
megathermal angiosperm family comprising
more than 500 species grouped into three
subfamilies with an intercontinental disjunct
distribution: Monotoideae in Africa, Madagascar,
and South America; Pakaraimaeoideae in
South America; and Dipterocarpoideae in the
Seychelles, India, and Southeast Asia ( 6 ). A
specific combination of morphological and eco-
logical characteristics determined the ecolog-
ical success of Dipterocarpoideae and enabled
them to out-shade the canopy of other tree
families, which led to their dominance in the
mature forests of Southeast Asia ( 6 ). These
characteristics include ectotrophic mycorrhi-
zal association, specific pollinators, mast fruit-
ing overcoming seed predation, protective resin
in multicellular secretory ducts, poorly nutri-
tious and resinous unpalatable leaves for many
herbivores, and plagiotropic followed by ortho-
tropic branching patterns in trees. Dipterocarps
have been the major source for timber extrac-
tion across Southeast Asia over the past
50 years, which has led to their overexploita-
tion and has left many formerly superabundant
species critically endangered.
We present eight fossil pollen types, extracted
using the standard palynological techniques ( 7 ),
referabletofivelivinggeneraofthesubfamily
Dipterocarpoideae (Dipterocarpus,Dryobalanops,
Shorea,Vateriopsis, andVatica) and one genus of
the subfamily Monotoideae (Monotes) (Fig.
1 and figs. S1 to S4) from the Maastrichtian
of Sudan and the Paleocene and early Eocene
of India (Fig. 2, appendix S1, and table S1).
These are combined with the secobicadinane
biomarkers of dipterocarps retrieved from the
Late Cretaceous sediments from central India
(fig. S5), phylogenetic analysis ( 8 – 13 ) of pollen
(fossil and extant) morphology, and DNA se-
quence data of 54% of the known Diptero-
carpaceae species to do the following: (i) tracethe origin and evolution of the family; (ii) de-
fine the role of climate and dispersal pathways,
including movement of the Indian Plate during
the Late Cretaceous–early Paleogene in the di-
versification of the family; and (iii) resolve the
paleobiogeographic history of lowland dipter-
ocarp rainforests in Southeast Asia.
The pollen of the subfamily Dipterocarpoideae
is distinctive in being tricolpate, with very long
colpi reaching almost to the poles, and with
exine consisting of a thin basal layer, except in
Vateriopsisand allied genera, with a much
thicker outer, sculptured layer. The columellae
and frequently grooved or crenelated ridge-
bearing tectum of the outer layer fuses into a
tilioid structure, except in the tribe Shoreae ( 14 ).
Pollen of most species of Dipterocarpoideae
are smaller than 35mm, whereasDipterocarpus
pollen ranges from ~50 to 100mm. The diag-
nostic characteristics ofDipterocarpuspollen
canmostlybeseeninlightmicroscopy(LM),
whereas the identification of other genera also
requires scanning electron microscopy (SEM)
examination. No other plant family exhibits
the combination of tricolpate configuration,
absence or reduced endexine, and tilioid exine
structure (table S2). By contrast, the subfamily
Monotoideae pollen is tricolporate with the
ektexine forming a fairly coarse tilioid struc-
ture ( 14 ). Phylogenetic analyses based on pollen
morphological characters confirm the place-
mentofsevenoftheeightfossiltaxawithin
five extant genera of Dipterocarpoideae, name-
lyVateriopsis,Dipterocarpus,Dryobalanops,
Vatica, and the three differentShoreasections—
Anthoshorea, Parashorea, and Rubroshorea
(figs. S6 to S8)—and the remaining one fossil
taxon was placed within the genusMonotes,
belonging to subfamily Monotoideae (figs.
S6 to S8). The fossil pollen types represent
five clades of Dipterocarpaceae based on our
molecular phylogenetic analyses and previ-
ously published phylogenetic literature ( 15 )
(Fig. 3B, table S1, and appendix S2).
The discovery of fossil pollen with clear
affinity to the subfamilies Dipterocarpoideae
and Monotoideae from the Maastrichtian of
Sudan and the Paleocene of India refutes fre-
quent references to the unreliability in de-
termination of fossil dipterocarp pollen ( 16 ).
The previous skepticism largely stems from
the fact that there are abundant records of
Dipterocarpoideae leaves and woods from the
Indian Neogene ( 17 ) but very few from the Late
Cretaceous and Paleogene. This disparity most
likely relates to many factors, including differ-
ences in the depositional setting. Other factors
to bear in mind are (i) the possibility that the
earliest phase of evolution of the family may have
involved a long period of mosaic evolution and
(ii) early macrofossils may not have borne ana-
tomical features recognizable as Dipterocarpaceae.
The presence of pollen comparable to that
ofMonotesin the late Paleocene and earlySCIENCEscience.org 28 JANUARY 2022•VOL 375 ISSUE 6579 455
(^1) Birbal Sahni Institute of Palaeosciences, Lucknow 226007,
India.^2 Palynova Ltd., Littleport, Cambridgeshire CB6 1PY,
UK.^3 Earth Sciences Department, Royal Holloway, University
of London, Egham, Surrey TW20 0EX, UK.^4 The Nature
Conservancy Centre, New Delhi 110024, India.^5 Department
of Earth Sciences, Indian Institute of Technology Bombay,
Mumbai 400076, India.^6 Department of Biology and Centre
for Structural and Functional Genomics, Concordia
University, Montreal, Quebec H4B 1R6, Canada.^7 Quebec
Centre for Biodiversity Science, Montreal, Quebec H3A 1B1,
Canada.^8 Petroleum Laboratories, Research & Studies (PLRS),
Ministry of Petroleum and Gas, Khartoum, Republic of Sudan.
(^9) Institut Français de Pondichéry, UMIFRE 21 CNRS-MAEE,
Puducherry 605001, India.^10 Organismic and Evolutionary
Biology, Harvard University, Cambridge, MA, USA.
*Corresponding author. Email: [email protected]
These authors contributed equally to this work.
RESEARCH | REPORTS